public static IDataTransform CreateForEntryPoint(IHostEnvironment env, ArgumentsForEntryPoint args, IDataView input)
{
Contracts.CheckValue(env, nameof(env));
var host = env.Register("Tree Featurizer Transform");
host.CheckValue(args, nameof(args));
host.CheckValue(input, nameof(input));
host.CheckUserArg(args.PredictorModel != null, nameof(args.PredictorModel), "Please specify a predictor model.");
IDataTransform xf;
using (var ch = host.Start("Create Tree Ensemble Scorer"))
{
var scorerArgs = new TreeEnsembleFeaturizerBindableMapper.Arguments()
{
Suffix = args.Suffix
};
var predictor = args.PredictorModel.Predictor;
ch.Trace("Prepare data");
RoleMappedData data = null;
args.PredictorModel.PrepareData(env, input, out data, out var predictor2);
ch.AssertValue(data);
ch.Assert(predictor == predictor2);
// Make sure that the given predictor has the correct number of input features.
if (predictor is CalibratedPredictorBase)
{
predictor = ((CalibratedPredictorBase)predictor).SubPredictor;
}
// Predictor should be a FastTreePredictionWrapper, which implements IValueMapper, so this should
// be non-null.
var vm = predictor as IValueMapper;
ch.CheckUserArg(vm != null, nameof(args.PredictorModel), "Predictor does not have compatible type");
if (data != null && vm.InputType.VectorSize != data.Schema.Feature.Type.VectorSize)
{
throw ch.ExceptUserArg(nameof(args.PredictorModel),
"Predictor expects {0} features, but data has {1} features",
vm.InputType.VectorSize, data.Schema.Feature.Type.VectorSize);
}
var bindable = new TreeEnsembleFeaturizerBindableMapper(env, scorerArgs, predictor);
var bound = bindable.Bind(env, data.Schema);
xf = new GenericScorer(env, scorerArgs, data.Data, bound, data.Schema);
ch.Done();
}
return(xf);
}
public static IDataTransform CreateForEntryPoint(IHostEnvironment env, ArgumentsForEntryPoint args, IDataView input)
{
Contracts.CheckValue(env, nameof(env));
var host = env.Register("Tree Featurizer Transform");
host.CheckValue(args, nameof(args));
host.CheckValue(input, nameof(input));
host.CheckUserArg(args.PredictorModel != null, nameof(args.PredictorModel), "Please specify a predictor model.");
using (var ch = host.Start("Create Tree Ensemble Scorer"))
{
var scorerArgs = new TreeEnsembleFeaturizerBindableMapper.Arguments()
{
Suffix = args.Suffix,
TreesColumnName = "Trees",
LeavesColumnName = "Leaves",
PathsColumnName = "Paths"
};
var predictor = args.PredictorModel.Predictor;
ch.Trace("Prepare data");
RoleMappedData data = null;
args.PredictorModel.PrepareData(env, input, out data, out var predictor2);
ch.AssertValue(data);
ch.Assert(data.Schema.Feature.HasValue);
ch.Assert(predictor == predictor2);
// Make sure that the given predictor has the correct number of input features.
if (predictor is CalibratedModelParametersBase <IPredictorProducing <float>, Calibrators.ICalibrator> calibratedModelParametersBase)
{
predictor = calibratedModelParametersBase.SubModel;
}
// Predictor should be a TreeEnsembleModelParameters, which implements IValueMapper, so this should
// be non-null.
var vm = predictor as IValueMapper;
ch.CheckUserArg(vm != null, nameof(args.PredictorModel), "Predictor does not have compatible type");
if (data != null && vm.InputType.GetVectorSize() != data.Schema.Feature.Value.Type.GetVectorSize())
{
throw ch.ExceptUserArg(nameof(args.PredictorModel),
"Predictor expects {0} features, but data has {1} features",
vm.InputType.GetVectorSize(), data.Schema.Feature.Value.Type.GetVectorSize());
}
ISchemaBindableMapper bindable = new TreeEnsembleFeaturizerBindableMapper(env, scorerArgs, predictor);
var bound = bindable.Bind(env, data.Schema);
return(new GenericScorer(env, scorerArgs, data.Data, bound, data.Schema));
}
}
